Data structures

Parse all image files

Check one object

Load an image

Visualize image with mask

U-Net demo

Data processing

Create Tensorflow Data Loader

Train/Validation Split

Training

The following training code and model code was adapted from

https://yann-leguilly.gitlab.io/post/2019-12-14-tensorflow-tfdata-segmentation/

Data Tuning Parameters

Model Saver Class

Custom Loss Function

In this case dice loss is used according to the implementation in:

https://lars76.github.io/2018/09/27/loss-functions-for-segmentation.html

Create and Compile Model

Create moodel saver instance

Training Hyperparameters

Start Training

Manually Save Model

Load

Process Prediction

Visualize Prediction